Friday, February 7, 2014

Lying with infographics and visualization

A few words about the most recent addition to my list of recommended infographics and visualization resources: Its title is Disinformation Visualization: How to lie with datavis, and it was written by Mushon Zer-Aviv (Twitter.) You should read it as soon as you can. Quoting:
"Our eyes are trained to process the visual environment as evidence, and to process the products of language as arguments. Therefore, whether beautiful or ugly, visual lies masquerade as evidence and slip under our radar. We, in return, readjust our worldview to accommodate these visual lies as we do when we witness evidence. As long as we do this, there will be those who will continue to tell these visual-lies, convincing us, and often themselves, that evidence will make itself evident and that subjective culture and political biases won’t creep in." 
Reading this piece was an eerie experience for me. A month and a half ago I turned in a chapter for a book that will be published by Springer this year. I discuss similar matters in it, albeit making a clearer distinction between lies —situations in which misinformation is the result of designers consciously building deceitful graphics— and mistakes —based on sloppy data analysis or design choices. The chapter will be freely available online as soon as the book is published for the reasons I explained here. This is its introduction:

Graphic Lies, Misleading Visuals

Can information graphics (infographics) and visualizations lie? Most designers and journalists I know would yell a rotund “yes” and rush to present us with examples of outrageously misleading charts and maps. Watchdog organizations such as Media Matters for America have recently began collecting them, and a few satirical websites have gained popularity criticizing them. Needless to say, they are all great fun.

The notion that graphics can indeed lie derives from Darrell Huff’s bestseller “How to Lie With Statistics”, which describes the most common kinds of visual sins, such as truncating the Y-axis of graphs. Edward Tufte, always keen on contriving catchy memes, tried to coat Huff’s teachings with a scientific-looking varnish and invented a unit of measurement called the ‘Lie Factor’. He even came up with a formula to calculate it. As it happens with much of Tufte’s oeuvre, it’s impossible to know if he was writing with tongue in cheek. I assume he did:

Lie Factor = Size of effect shown in the graphic / Size of effect in data

(The closer the Lie Factor is to 1.0, the more accurate the graphic is.)

Playful quantifying efforts aside, let me take the risk of sounding platitudinous: Charts, graphs, maps, and diagrams don’t lie. People who design graphics do. This is a no-brainer if we stick to a well-known definition of the word “lie” in the literature about ethics: “An intentionally deceptive message in the form of a statement” (Bok, 1999). The graphic is that statement, not the agent who creates it and delivers it.

A graphic can mislead, though. Misleading is not the same as lying because a graphic can lead readers astray without the conscious intervention of its designer. This distinction is not a technicality. According to professional ethics codes in journalism and graphic design, knowing the truth and hiding it, or conveying it in a way that distorts it is simply unacceptable.

Codes of conduct are based on a priori rules, duties, and obligations. They are the embodiment of what’s called deontological ethics in the academic literature. Therefore, according to them the intention of the agent is the key to analyzing if the actions she performs are right or wrong. On the other hand, designing a misleading graphic as a result of naive mistakes while analyzing or representing data is ethically neutral. I guess that we all can agree that lying is the worst action any visual communication professional can perform, so lying will be the focus of the first part of this essay.

However, I find the alleged ethical distinction between lies and mistakes intriguing and worth playing with. Let’s ask ourselves: Are the intentions of the designer really enough to evaluate the appropriateness of each graphic she creates? My hunch is that they aren’t. Let me elaborate.

Professional codes of conduct implicitly differentiate between truth and truth-telling. In many circumstances, truth may be unattainable, fuzzy, or even unknowable, but that doesn’t spare us of our obligation of being truthful. Someone who intends to communicate a message may not know all information needed or — more importantly for my argumentation — may not possess the skills to represent it correctly. But if she strives to do her best, she will be acting ethically.

So far, so good. This is simpler than the stuff discussed in any Ethics 101 course. As a designer or journalist, you are likely feeling comforted at this point: “Perhaps I make mistakes sometimes but I certainly don’t lie,” you may be thinking. Allow me to distress you a bit. Don’t think just about yourself, the designer, when evaluating your own decisions while creating a graphic. Think of who’s going to decode it, too. Read the following paragraph carefully:
When we see a chart or diagram, we generally interpret its appearance as a sincere desire on the part of the author to inform. In the face of this sincerity, the misuse of graphical material is a perversion of communication, equivalent to putting up a detour sign that leads to an abyss (Wainer, 2000).
Notice that Wainer doesn’t imply that the author is trying to actively deceive you. The “misuse of graphical material” could have been completely unintentional, the same way that putting the detour sign in the wrong side of the road could be the product of sloppiness or ignorance.

Switching the focus back and forth from the designer (the encoder) to the reader (the decoder) makes the distinction between lying and misleading much blurrier, for in the world out there the consequences of both lies and mistakes are equally grave: More noise and misinformed audiences. And so, paradoxically, I feel that what was going to be a short piece to help myself reflect about lying with information graphics cannot stick to lies alone. If we agree that infographics must represent a reality — data, information — with accuracy, we mustn’t just obsess over the conscious actions of communicators. We should also point out the responsibility we have to educate ourselves to overcome our own biases, shortcomings, and knowledge gaps. We must work hard to eliminate or, at least, to minimize ambiguity, confusion, and potential errors of interpretation in our graphics. That will be the core idea of the second part and the conclusion of this essay.